One deep discharge. One empty-to-full charge window. From a unit with a clearer low-to-full gap than our denser sample.
This TMT250 ran from about 90% down to 3% over roughly eight days of ordinary use — a long drain you can follow in the retained pings, not a single short ride.
At 04:35 UTC the pack was at 3% (3418 mV). By 07:06 it reported 100% (4154 mV). Charging finished inside that ~2.5 hour window.
Source: prod AVL for IMEI …884900. Chosen over denser …840056 because the empty-to-full gap is observed within ~2.5 h instead of an overnight hole. Discharge is the multi-day decline to 3%.
Pings are sparse — not a fleet lifetime study. Charge duration is an upper bound between observed points. Prod Mongo currently retains little TMT history beyond a few IMEIs.